ViewR 2.0.0

Major new architecture: the viewdt() data explorer widget

ViewR 2.0.0 introduces an entirely new, dependency-free htmlwidget data explorer built from scratch in vanilla JavaScript (no React, no build toolchain). It is portable: render it in the RStudio/Positron Viewer, inside Shiny, in R Markdown / Quarto, or export it to a standalone HTML file.

• viewdt(data, options = viewdt_options()) — opens the modern explorer.
• viewdt_options() — configure theme, labels, hidden columns, NA string, visual features, and the viewer-cap safeguard.
• save_viewdt() — export any data frame to a portable, offline HTML grid.

Explorer features

• Virtualized grid that paints only visible cells, with sticky headers and a sticky row-index column; click a row to pin it.
• Kaggle-style micro-dashboard headers: data-type badges, mini distribution spark-histograms (numeric) / top-category stacked bars (categorical), and a colour-coded data-completeness (missingness) bar.
• Column metadata cards on the info icon: rows, unique, missing %, and Min/Mean/Median/Max or Top-5 categories.
• Data Insights drawer with an interactive SVG histogram (numeric) or Pareto bar chart (categorical), completeness meter, and descriptive stats.
• Visual query builder: multi-condition AND/OR filters with type-aware operators and a searchable multi-select for categorical columns.
• Column visibility picker and global search.
• Reproducible code generator: copy-pasteable dplyr, base R, and SQL matching the active filter + visible-column state.
• Light / dark / auto theme, variable-label display, and custom NA string.

The legacy Shiny-gadget editor remains available as ViewR().

ViewR 0.2.0

New features

• Excel-like editor (edit = TRUE): powered by rhandsontable, supports inline cell editing, row addition via right-click context menu or the Add Row button, and unlimited undo / redo.

• Find & Replace tab: search for a literal string, regex pattern, or exact cell value across one or all columns. A live Preview diff table shows matched cells before changes are committed.

• Variable Info tab: one-row-per-column summary of data type, N, missing count, missing %, unique count, min, max, and sample values.

• R Code tab (generate_code = TRUE): live dplyr pipeline that reflects all current filters, sorts, column selections, and find-and-replace operations. Copy to clipboard button included.

• Multi-condition Filters with AND / OR logic and eleven operators: ==, !=, >, >=, <, <=, contains, starts with, ends with, is NA, is not NA.

• Multi-column Sort: add as many ordered sort levels as needed, each independently ascending or descending.

• Column visibility panel: show / hide any column; bulk All / None toggles.

• Variable labels: automatically read from column "label" attributes (e.g. from haven::read_spss()); displayed as hoverable tooltips on column headers in the Data View table.

• Twelve Bootstrap themes via shinythemes: "flatly", "cerulean", "cosmo", "darkly", "lumen", "paper", "readable", "sandstone", "simplex", "spacelab", "united", "yeti".

• install_viewr_deps(): helper to detect and install all required packages in one call.

• Pipe-friendly: ViewR() accepts piped input and returns the working data frame invisibly when the user clicks Done.

Bug fixes

• Type coercion after find-and-replace no longer depends on methods::as(); replaced with an explicit switch() over the original column class.

Internal changes

• Separated UI, server, helpers, and code-generation concerns into dedicated source files (R/ui.R, R/server.R, R/helpers.R, R/code_gen.R).
• Full testthat (edition 3) test suite covering filter engine, sort engine, find-and-replace, variable info, and code generation.
• Added vignettes/ViewR-intro.Rmd introduction vignette.
• Added _pkgdown.yml for the package website.
• Hex sticker added to inst/sticker/.

ViewR 0.1.0

• Initial release: basic ViewR() stub.